Love, Money, and HIV

One

A Stubborn Disparity

The bustling city of Kisumu, the capital of Nyanza province, Kenya, lies nestled by Lake Victoria, the largest freshwater lake in Africa, and only a few miles from the equator. A stroll through its main streets yields a display of an apparent clash of worlds and cultures. There is the Africa we know: the busy human drama of mitumba (secondhand clothing) women, market traders and hawkers, street children, fishermen and the pungent smell of fresh and smoking fish, house flies and mosquitos, numerous bicycles and matatus (public transportation vehicles) hooting through the town, and travelers stopping by on the way to or from Uganda or Tanzania. And there is the Africa we are coming to know: tarmaced streets, a prominent Citibank building (taken up by another company), Internet cafes, international nongovernmental organizations (NGOs), and English billboards prominently advertising pay-as-you-go Safaricom mobile phones and Trust condoms. The exuberance, color, and activity of Kisumu, however, mask the slow and terrible unfolding of a demographic catastrophe. A survey conducted in the late 1990s revealed that, among its young women, almost 30% of 15-19 year olds and almost 40% of 20-24 year olds were HIV positive, carrying a virus that would kill them in six to ten years. These were three to six times the HIV rates of same-aged men.1 Unfortunately, the gender disparities in this survey were not unique to Kisumu, Kenya, but were reflected in study after study across sub-Saharan Africa, the world's most affected region. Despite having only 12% of the world's population, it has 69% (23.5 million) of people living with HIV/AIDS, 70% (1.2 million) of AIDS-related deaths, and 71% (1.8 million) of new infections.2

Women now make up approximately 60% of Africans living with HIV/AIDS,3 and young women are at particular risk. In 2001, in AIDS, the official journal of the International AIDS Society, an influential editorial comment was published with the title, "To Stem HIV in Africa, Prevent Transmission to Young Women." The authors noted that "the high HIV prevalence among women aged 15-19 years could be critical in provoking and maintaining an explosive HIV epidemic."4 The editorial concluded by urging policy makers to focus their efforts on this group of young people. The issue becomes much more salient in light of the current burgeoning youth population in sub-Saharan Africa: 41% of Africans are under age 15.5 Girls in this demographic group are on the brink of their greatest period of risk of contracting HIV. This situation offers both an extraordinary opportunity to halt the epidemic and potential disaster if the HIV epidemic is not stemmed.6

Despite the early call for action, gender differences in HIV rates among youth in sub-Saharan Africa continue to be widespread and have been confirmed over the subsequent decade in several larger and more representative surveys than those that prompted the comment.Further, these disparities tend to persist until youth reach their mid-30s; after that point, we start to see substantial variation by country in which gender has higher HIV rates.7

Figure 1, for example, shows findings from a sample of national surveys of HIV-prevalence rates among young men and women in selected high-prevalence countries in sub-Saharan Africa. It illustrates not only the high burden of HIV in countries such as South Africa, which holds the world's largest HIV-positive population (over five million people), but also the variability across countries in the size of the disparity between young men and women. In South Africa, 31% of 21-year-old women are HIV positive, a rate that is five times higher than same-aged men, who have a prevalence rate of 5.6%. However, in Zimbabwe, 15-19-year-old females' 6.2% HIV prevalence is only twice as high as same-aged men, who have a prevalence rate of 3.1%. On average, young African women are three times more likely than young men to have HIV;8 however, as the figure illustrates, this average masks widely varying levels of disproportionate risk. Indeed, this variability in youth-gender disparities across the region suggests that this is not merely a story of biological sex differences. Something more is at work.

A Life-Course Approach to HIV Risk

The youth-gender disparity from ages 15 to 24 matters because young women have so much more to lose than men in years of life that are disease free and in life expectancy. As figure 1 illustrates, substantial numbers of people are contracting HIV on the cusp of adulthood. In the absence of antiretroviral medication, many are dying early in their marriages, leaving behind young children. For men, not only are they acquiring HIV at lower rates than women in their adolescence and early adulthood, but their peak HIV prevalence is also much later, in their late 30s or early 40s. This means that they lose far fewer years relative to their life expectancy compared to women.9 An important place to start, then, in investigating the gender disparity in HIV rates is to investigate the period in the life course when these disparities emerge and self-perpetuate-the transition to adulthood.

A "transition to adulthood" framework provides a lens through which to examine social processes and experiences salient to youth in the period between puberty and adulthood. It is a useful way of exploring the lives and developmental changes and challenges experienced by African youth by allowing the examination of key transitions such as relationship formation processes (including sexual initiation and the culmination of relationships in marriage or a stable partnership), the pursuit of education, finding employment, attaining financial independence, becoming a parent, and transitioning residences (principally moving out of the parental home).10 Importantly, these life course transitions co-occur: relationship decisions can be shaped by educational decisions; financial independence can determine residential transition.11 This co-occurrence of many life-changing and life-impacting decisions creates for young people a thick web of life events that is reflected in the fact that many experience their transition to adulthood as complicated and confusing. Because of the lifelong significance of many decisions made during this period, transitions to adulthood are sometimes characterized as risky. This recognizes the fact that mistimed (early or delayed), missed, or unsuccessful transitions to "normal" adulthood, or events and actions that interrupt an otherwise smooth transition, can have potentially long-term implications for educational attainment, lifetime income, and job prospects and health. Substance abuse, juvenile delinquency, teenage pregnancy, and HIV, for example, can complicate or even derail a young person's successful transition to adulthood by limiting their ability to enact other transitions such as finding employment, achieving financial independence, and forming stable long-term relationships.

Additionally, it is important to acknowledge that transitions have a processual nature. In other words, the establishing of a stable long-term relationship (and even parenthood), the completion of education, and stable employment are all transitions that can be begun, ended, elongated, interrupted, restarted, and returned to.12 My interest in a life-course approach, and the transition-to-adulthood perspective in particular, is to examine not so much the point at which adulthood is reached (though this is touched on), but rather the process of its attainment. This facilitates an analytical strategy that uses a focus on various transitions to adulthood to bring into relief the ways in which gender disparities in HIV rates emerge and are produced among young people.

The Book

The consistently higher HIV rates of young women across a subcontinent with diverse countries, cultures, and ethnic groups belie a phenomenon that is only captured by biological sex differences or a patchwork of localized stories. Rather, it suggests that overarching social-structural processes are also occurring across the continent, mapping onto biology, individual decision-making, and culture in local spaces. The combination of these processes places young African women transitioning to adulthood at great risk, while producing temporary safety for young men. At the heart of this book is a study of how one of these processes-consumption, the desire for and purchasing of modern goods-has come to play a crucial role in producing gendered life and death outcomes among young people.

The startling statistics cited at the beginning of this chapter struck a deep personal chord as a Kenyan native with a paternal grandmother from Nyanza province, who I regularly visited during school holidays. These were not distant statistics. They numerically represented a demographic category I inhabited. The age group, gender, and ethnic heritage of those at the highest risk at the time of the survey perfectly characterized me as well as several female cousins living in or near Kisumu, Nyanza's capital. Further investigation of the statistics revealed that young African women were at greater risk compared not just to men in Nyanza province, but to men in almost every African country for which I could find data. Why were young African women at much greater risk for HIV compared to same-aged young men? Biology was clearly part of the explanation, but the variation in the disparity from setting to setting suggested that something more was going on, something that I felt survey analysis alone could not quite capture. I designed a mixed-methods study to explore this disparity. Drawing on a life-course framework, the study explored young people's transitions to adulthood in the context of an ongoing HIV epidemic, drawing on the perspectives of 185 young people, middle-aged adults, and older adults, as well as 20 key informants. Rather than abandoning quantitative data, I chose to move back and forth between population-based survey data from Kenya, ethnographic and interview-based fieldwork in Nyanza province, Kenya (the setting that launched my interest in this topic), and published quantitative and qualitative analyses from several parts of sub-Saharan Africa in order to come to a more complete understanding not only of the plight of young women in Nyanza, but also of those living in many other parts of the continent.

This book examines how young African women navigate their relationships, schooling, employment, and financial access in the context of a devastating HIV epidemic and economic inequality, where extreme wealth is increasing alongside extreme poverty. Billboards in rural and urban sub-Saharan Africa advertising modern lifestyles and consumer products are an indelible part of the social landscape. This means that modernity and the consumptive goods and practices that signify it are present and highly visible in many rapidly globalizing parts of the continent, but just out of reach for many.13 These developments are deeply gendered and have implications for young women's HIV rates. The book examines the compounding of young women's desire for consumer products that require continual replenishment with the gendered and generational nature of access to income as well as resources. Many young African women are situated within stages in their life course and social structures where access to money, resources, and paid work is tightly constrained. Continual consumption requires partners with continual access to income. This ultimately makes intimate relationships with older, employed men, who have higher HIV-prevalence rates, more attractive than those with unemployed young men, who have relatively low HIV rates.

Let us briefly listen in on a conversation typical of the many I had with groups of high school girls in Nyanza. We had been talking about who they most saw affected by HIV/AIDS, and they said it was young women. I asked them why, and the following unfolded:

Sanyu: Why do you think it's mostly girls? Why not boys?

Anyango: Because girls are being attracted to so many things.

Mary: They are easily swayed by money.

Sacha: This is because of the financial status of the family, whereby if the family is poor, you know as girls we need several things, but if the family can't provide all this, the girl will be forced to search for them somewhere else, and if you find somebody who cannot provide all these necessities, the girl may switch to the next man she thinks might provide it. So that's the reason as to why the girls move from one man to another, and get infected to this disease easier.

Sanyu: So what things do girls need? It's been a while since I was in high school.

Jane: Cosmetics, maybe your friends look smarter than you . . .

Rosa: Good dresses.

Anyango: And she doesn't have . . .

Sacha: Which will force you to search for a man who can provide all this. [Laughter.]

Sanyu: So that means it can't be a fellow guy in high school? 'Cause he doesn't have the money? Or does it not matter? [Silence.]

Rosa: It must just be somebody from, someone who is financially stable, who can provide for all this.

Sanyu: And what age would that be . . . someone not in school?

Jane: No, those in the working class.

In thinking through this and many other such conversations, a number of things became apparent. First, their knowledge about the epidemic was not lacking. They were aware that their demographic-young women-had the most risk for HIV, and further, they were aware that they could get HIV from sexual relationships, with the most risk coming from having multiple partners. Second, they were clear that the reason why girls were most at risk was because of their desire for nonsurvival consumption, and that it was men who worked-"working-class" men-who were most able to provide them with the money to consume. The fact that things like cosmetics and good dresses were framed as necessities or needs led to a compulsion, indeed something that would "force" them to look for a man to help.

Many elements of this conversation are familiar to Africanist and AIDS scholars, who have similarly observed and written about transactional sex relationships, intimate relationships where money and gifts are exchanged but in which issues of love and trust are also considered at stake. However, few have moved beyond discussing the nature of these relationships to unpacking why young African women want to consume in the first place. In other words, while much is known about the existence and widespread nature of these relationships, little is known about why and importantly how young women across the continent have historically come to desire consumption to this degree, despite widespread knowledge about HIV/AIDS and widespread experience of the emotional and social devastation that has accompanied it. In using a mixed-methods approach to examining how girls transform into "consuming women" as they transition into adulthood, this book will not only unpack these puzzles, but also, in the process, shed light on three startling quantitative paradoxes in the AIDS literature that are part of the problem of young women's disproportionate HIV risk: why women living in the wealthiest households in Africa have the highest rates of HIV (chapters 4 and 5), why the least-educated African women have the lowest rates of HIV (chapter 5), and why African women who work for pay have higher rates of HIV compared to those who do not (chapter 6). Bringing to the fore the voices, thoughts, and experiences of young African women and men in the chapters that follow, along with representative survey data, this book will show how the complicated entanglement of love, money, and young women's transformation into "consuming women" lies at the heart of their disproportionate HIV rates.

The book examines how desires for money, gifts, modernity, and consumption become both gendered-with girls constructed as desiring more than boys-and inextricably linked with girls' intimate relationships with the riskiest male partners. Specifically, I illustrate how consuming young women have been cultivated and produced in three contexts-communities, schools, and labor markets. I explore the historical and contemporary construction of "gendered needs," young women's continual and disproportionately greater "needs" for modern products that must be bought, such as sanitary pads, cosmetics, and clothing compared to young men. I then situate these desires in three social-structural contexts: in communities, where the means of achieving beauty and attracting partners have over time become commodified (chapters 3 and 4); in schools, which are settings charged with the production of modern subjects, a modernity that young women's consuming practices signify (chapter 5); and in gendered labor markets, where the gap between young women's "needs" and their access to money, resources, and paid work becomes particularly acute, and where transactional relationships take on new significance (chapter 6). In the case of gendered labor markets, I further show how changing ecological environments mapped onto gendered economies can shape the sexual economies in which transactional relationships unfold with devastating effects on young women's HIV risk.

In the rest of this chapter, I first review the literature on gender differences in youth HIV rates, and then discuss the setting, data, and methods used in the study, before concluding with a brief overview of the chapters in this book.

Gender Differences in Youth HIV Rates

An extensive literature has emerged to examine factors that might explain why young women have higher HIV rates than young men. These factors can be loosely grouped into three sets of explanations: biophysiological, proximate, and social structural. However, many of these explanations not only are limited in their ability to account for the stubborn gender disparity in youth HIV rates across Africa's varied settings, but also suggest particular vulnerability not just for young women, but also for young men. A secondary aim of this section is to introduce readers who may be less familiar with the HIV/AIDS literature to some of the key factors underlying its spread.

Biophysiological Explanations

A particularly compelling explanation for the disparity in HIV risk between young men and women is young women's higher biophysiological susceptibility to the virus.14 The HIV virus is contracted through exposure to four transmitting fluids: semen, vaginal fluid, blood, and breast milk, with the highest concentrations of the HIV virus occurring in blood and semen. In sub-Saharan Africa, most young women's HIV risk is the result of exposure to HIV through heterosexual intercourse.15 Even if a young woman has only one sexual partner, there are several biophysiological reasons why she might be at greater risk of contracting HIV compared to a young man. For example, there is a higher viral load in semen compared to vaginal secretions. Additionally, young women's bodies are particularly vulnerable when they are still developing, and they thus are more likely to experience genital trauma during sex that may create more openings for the virus to enter. The presence of blood when virginity is lost is also likely to place young women at higher risk of contracting HIV compared to young men.

However, findings on sex differences in transmission probabilities of HIV are mixed. While many studies from Europe and the United States suggest greater male-female transmission probabilities (that is, a man is more likely to transmit HIV to a woman), results from developing countries such as Thailand, Kenya, and Uganda suggest the opposite trend. These studies suggest that women have higher transmission probabilities than men and are more likely to transmit HIV to men than vice versa. Part of the explanation for this is that men in developing countries are more likely to have high-risk partners such as commercial sex workers. Because commercial sex workers have many sexual partners, they are a population that is particularly vulnerable to being exposed to HIV, as well as to passing HIV on to others. Sex workers also have a higher likelihood of being exposed to other sexually transmitted infections (STIs).16 Having an STI increases the likelihood of contracting HIV because STIs create additional openings for HIV to enter and may also increase the amount of HIV in vaginal fluids, thus increasing women's transmission probabilities.17

Uncircumcised men also experience higher HIV risk compared to men who are circumcised. Many of the countries and ethnic groups greatly affected by the HIV epidemic have lower male circumcision rates.18 In the most thorough confirmation of this finding, researchers enrolled several hundred men in Kenya, South Africa, and Uganda, randomly divided them into two groups, and circumcised one group, while leaving the other group intact. Following them over several months, they found significantly lower HIV rates among the men who were circumcised in all three countries. In Kenya, for example, men from Kisumu, the town profiled at the beginning of the chapter, who were circumcised experienced a 53% reduced risk of acquiring HIV compared to the men who had been left intact. Biological studies also suggest that circumcised men face reduced risk for the acquisition of ulcerative STIs, which appear to greatly exacerbate the risk of acquiring HIV.19 Transmission probabilities, STI acquisition, and the lack of circumcision are all factors that suggest vulnerability for young African men. In sum, while biophysiological factors clearly play a role, they are not in and of themselves sufficient explanations for young women's consistently higher prevalence levels across sub-Saharan Africa. Further, biology does not explain the variance in gender differences in HIV rates. Why, for example, were Zimbabwean women twice as vulnerable as their male counterparts, while South African women were five times as vulnerable, as shown in figure 1?

Proximate Explanations

Since heterosexual intercourse is a central mode of HIV transmission in Africa, perhaps how different cultures and ethnicities shape the proximate determinants of the sexual acquisition of HIV is the most important factor.20 That is, factors that directly affect (1) how much infected sexual fluid an individual is likely to be exposed to, and (2) how risky that exposure is. Exposure to infected fluid can be examined through factors such as age at first sex (affecting length of sexual activity) and number of sexual partners (and partner's partners). The risk of exposure can be examined through factors such as frequency of condom use (the most effective means of preventing sexual HIV transmission).

Age at First Sex

A particularly consequential factor that might be affected by culture is the age at which an individual begins sex. In addition to the presence of blood at first sex, an early sexual debut also affects the length of time an individual has been sexually active and potentially exposed to HIV. For example, in Nyanza province, while men's median age at first sex is 17.4 years, women begin sex almost half a year earlier at 16.7 years.21 This means that women experience a longer duration of potential exposure to HIV compared to men, which may partially underlie their higher HIV rates. However, in other parts of Africa with high HIV prevalence, this factor seems less important. In Manicaland, Zimbabwe, for example, the median age at sexual debut was the same for men and women (18.5 years) despite the 20%-25% difference in HIV prevalence rates. In South Africa, men begin sex earlier than women, despite young women's having five times the HIV rates of same-aged young men.22

Age and Number of Sexual Partners

The more sexual partners an individual has, the greater theirchance of encountering an HIV positive partner. This would also be a compelling explanation for young women's higher rates except for the fact that young men typically report more sexual partners than young women.23 Women's primary risk occurs because their fewer reported partners are typically older, and in some cases considerably older than they are. While young women have higher HIV rates than same-aged men, in many settings they have lower HIV rates than men five to ten years older. Thus a 17-year-old female with only one 25-year-old sexual partner is more likely to be exposed to HIV than another same-aged female with three 17-year-old sexual partners.24 However, one could also make the opposite argument. Young men's sexual partners are highly risky partners; their female counterparts aged 15 and above have higher HIV rates than them. Thus young men's likelihood of encountering an HIV-infected same-aged partner would actually be higher than that of a young woman. So to the extent that a young man is sexually active, his same-aged partners are of far greater risk to him than he is to her.

Condom Use

These relationships would not be as risky if every sexual act was protected by condom use, the most effective method of prevention of sexually transmitted HIV (aside from abstinence). While condom use has increased significantly over the past few years and knowledge of condoms was over 85% in several sub-Saharan African countries, in most of these countries only about a third to a half of sexually active young men aged 15-29 reported condom use at last sex. Older and especially married men's condom use is dramatically lower, as is self-reported use among young women. The abiding problem for many public health officials working on HIV interventions in sub-Saharan Africa is the continued resistance to widespread and consistent condom use before, within, and out of marriage.25

Connective Social-Structural Explanations

These are some of the major proximate explanations underlying young women's higher HIV risk, and they are useful to varying degrees in understanding the phenomenon in particular cultures, settings, and countries. However, they are incomplete in explaining why young women's greater HIV risk is replicated across so many diverse settings in sub-Saharan Africa despite different combinations of factors, some of which seem to clearly disadvantage young men more than young women. It does not seem to matter what unique configuration of these factors are operating in a given locale-young women still have higher HIV rates compared to young men.

This stubborn gender disparity in youth HIV rates suggests that there are common overarching social-structural processes at work across a number of settings that are connecting multi-level factors together in similar ways. Connective social-structural processes organize biophysiological and proximate determinants of HIV risk in similar ways, and thus can account for similar gendered and generational outcomes across multiple geographic spaces and cultures. Two of the most prominent connective social-structural processes discussed in the literature are sexual networks and migration.

Sexual Networks

The study of sexual networks involves capturing a bird's-eye view of structural patterns of sexual relationships in particular communities, enabling an examination not only of an individual's sexual relationships, but also of the relationships of their sexual partners' partners, ultimately revealing how individuals are connected to other individuals through sexual relationships. Thus studies have mapped entire romantic and sexual networks within a US high school and on an island in Malawi, for example.26 Such studies make visible the HIV risk a young woman may have even if she is monogamous. For example, in figure 2, Liz only has a single sexual partner, Tom. However, Tom is also having concurrent sexual partnerships with Jane and Crystal. Crystal is HIV positive and transmitted HIV to her husband, Jim. In this sexual network, then, Liz is at great risk of HIV acquisition because her sexual partner's partner Crystal is HIV positive. If Tom continues his relationship with Crystal without protection, over time, he will get HIV, which he might then pass on to Liz as well as Jane. In this way, HIV can spread through a sexual network, affecting many people in a short space of time. 

In this case, Liz may not be aware that Tom is being unfaithful and may not think she is at risk. However, in many settings across the continent, men's concurrent partnerships are normalized, culturally accepted and even institutionalized.27The most common institutionalized form of concurrency is polygamy where men can legally (as recognized by the state through customary law) have more than one wife. Outside of or in addition to this, informal arrangements ("outside" wives or mistresses) are also common.28 Thus one HIV infected man can pass HIV on to his wives and extramarital or nonmarital partners ultimately contributing to larger numbers of infected women relative to men.29 These features of sexual networks appear in several settings across the continent.

However, women can also place men at risk through their own concurrent partnerships. A study analyzing population-based representative surveys with HIV sero-discordant couples (where one member is HIV positive and the other HIV negative) found that in 30%-40% of cases women were the HIV-positive partners.30 This suggests that the case of Crystal in figure 2 is not rare. However, when we take a life-course perspective to HIV risk, young women like Liz may be acquiring HIV in a premarital relationship, and then getting married to a man with more than one wife, thus introducing HIV into the sexual network. In these ways, then, women's pre- or extramarital relationships with high-risk partners (such as older men) or multiple partners can also place their marital partners at high risk of acquiring HIV. Additionally, from a cross-sectional perspective, more men than women may be involved in concurrent partnerships overall, thus putting multiple women at risk for HIV at a single point in time; however, from a life-course or longitudinal perspective, young women's concurrency, such as that of Crystal, both during premarital relationships and entering into polygamous first marriages as junior wives may also be placing multiple men at risk for HIV.31

Migration

Migration and mobility are also key connective social-structural processes that have facilitated the rapid spread of HIV across large spaces in Africa. Migration essentially puts sexual networks in motion. Much of the Southern African epidemic, it is argued, is rooted in colonial-era labor migration systems still in place today, which separate workers from their families with only periodic returns home.32 At either end, opportunities for concurrent relationships exist for both men and women. Male workers living in single-sex residential environments seek sexual relief among commercial sex workers or engage in longer-term relationships with women in their work environments. Meanwhile at home, wives and partners might stray for loneliness, or in some cases they might seek economic support if their husbands do not send money home. Unprotected partnerships at each end provide ample opportunity for HIV infection to be acquired and then passed on to a partner during reunions.

Along with this, HIV-prevalence rates are particularly high among other kinds of people who are occupationally mobile. Commercial truck drivers who transport goods along major trade routes across Africa, for example, have been found to have particularly high HIV-prevalence rates and to some extent serve as vectors for the disease, engaging in relationships with women at different truck stops along Trans-African highways.33 However, these women might also serve as bridges of disease transmission between high-risk truckers and the local populations in which they reside, an important sexual-network link in epidemics.34 In addition to the highways, several of the areas hardest and earliest hit by HIV in Kenya (Nyanza province), Uganda (Rakai district), and Tanzania (Mwanza province) lie around Lake Victoria, Africa's largest freshwater lake and the source of the Nile, another major space across which goods are transported and traded. Fishing communities-men and women alike-have been particularly devastated by HIV.35

Other potential vectors of the disease are large-scale movements of armies, rebels, and refugees as a result of regional conflict,36 as well as the voluntary, commercial, and coercive sexual encounters that often follow. HIV transmission in these settings is fueled by the widespread presence of STIs (as noted earlier, an important cofactor of HIV) reflecting poor public sexual-health systems, which often break down in times of war.37 Young men seeking economic opportunity in mines or major towns, as well as young soldiers and rebels, may engage in high-risk behavior while away from their wives and partners. Many young women may be dependent on remittances or relationships with resident men for access to income, or they may be vulnerable to sexual violence as refugees or victims of conflict. These factors, in combination, imply risk for both young men and women.

In this book, I focus on another connective social-structural process, one that emerged as key to more clearly explaining gender differences in HIV outcomes among young people-consumption. Specifically, I show (1) how consumption desires and practices differentially shaped young women's and men's individual sexual decision-making, thus directly affecting their proximate determinants of HIV risk, (2) how consumption operated at a social-structural level within institutions such as schools and labor markets to shape the environments and structure of the sexual networks in which young women in particular were making these decisions, and finally (3) the similarity of these gendered and sexual economy dynamics across several high-prevalence settings throughout sub-Saharan Africa. In the remaining portion of this chapter, I discuss the setting, data, and methodology used in this study, and provide a brief overview of the book's chapters.

Study Setting

Kenya is an East African country, 582,646 square kilometers large, that lies along the Indian Ocean, and is bordered by South Sudan, Somalia, and Ethiopia to its north, Uganda to its West, and Tanzania to its South. A former British colony, it gained independence in 1963, led by its first president, Jomo Kenyatta. Kenya has had three presidents since then; the most recent, Uhuru Kenyatta, the son of the first president, was elected in March 2013. Kenya is considered a low-income country. It ranks 145th out of 186 on the human development index; it has an annual GDP of $33.6 billion and a GNI of $1541 per capita. Almost half the country (46%) lives on or under the poverty line ($1.25 a day). About 24% of the population lives in urban areas, and about 87% of the adult population are literate. In the census from 2010, the country's population was estimated at 39.4 million people. The country has a 2.7% annual rate of population growth with a total fertility rate of 4.6 per woman.38

Kenya is divided into eight provinces. Nyanza, the site of this study, which is located in the West, is the third largest, with 5.4 million residents, about 14% of the country's population. The Luo are a Nilotic ethnic group primarily living in Nyanza. They are the third largest ethnic group in the country, following the Kikuyu and the Luhya ethnic groups. While several studies have been undertaken in this area over the past decade, and while a lot of money has been spent, many communication campaigns have been mounted, and several nongovernmental organizations (NGOs) have invested in fighting HIV in Nyanza, the epidemic persists.39 Why this is the case remains a mystery. Nyanza also provides an example of a province with macroeconomic elements common to other parts of Africa that are affected by HIV: it is a labor migrating province, situated along major trading routes, including the Trans-African highway. However, despite large volumes of trade passing through Nyanza, many of its inhabitants remain poor. District Development Plans of the districts in this study recorded absolute poverty rates (proportions of people living on less than $1 a day) ranging from 53% to 69%, making it one of the poorest provinces in the country. In Kenya as a whole, 8% of women and 4.3% of men were HIV positive. However, as figure 3 illustrates, Nyanza province has the highest rates in the country. One in six women (16%) and one in ten men (11.4%) are HIV positive. The Luo ethnic group also has the highest HIV rates in the province,40 as well as in the country as a whole. One in five Luo women (22.8%) and one in six Luo men (17.1%) are HIV positive (figure 4).41

Studying the epidemic among the Luo in Nyanza province is particularly significant because it is in the region where the HIV epidemic began and that has borne the HIV/AIDS epidemic the longest. After HIV became an epidemic in Central Africa in the late 1970s, the disproportionate early burden of the epidemic in sub-Saharan Africa was in the region around Lake Victoria, Africa's largest freshwater lake and the source of the River Nile. Many of the hardest hit locations in the 1980s and 1990s were in lakeside districts and provinces in Uganda (Rakai district), Tanzania (Mwanza and Bukoba provinces), and Kenya (Nyanza province), where the first cases of HIV/AIDS were reported. The epidemic then spread across Eastern Africa before proceeding to Southern Africa, which currently bears the heaviest burden.42 Understanding the historical and current HIV epidemic dynamics in this setting may yield significant and fresh insight into settings where the HIV epidemic has now become endemic and seemingly impervious to HIV prevention efforts.

Social and Ecological Setting

The most dominant geographic feature of Nyanza province is Lake Victoria. The widespread dependence of many Africans on the ecological and natural environment for their livelihood (for food, for water, for income) both in this setting and in many other parts of the continent makes it an important though often ignored actor in accounts of HIV epidemics.43 I will show how the environment was consequential in shaping young women's HIV risk in chapter 6.

I focused my fieldwork in four Nyanza province districts that border Lake Victoria: Bondo (north of the lake), Kisumu (central, and which includes the capital city), Nyando (central), and Homa Bay (south of the lake). Following the ratification of the constitution in 2010, Kenya adopted a county system, transforming 72 districts into 47 counties. With this system, these four districts are now subsumed under three counties: Siaya county (Bondo district), Kisumu county (Kisumu and Nyando districts), and Homa Bay county (Homa Bay district). These settings were selected to provide geographic diversity, while remaining in what is known as Luo-Nyanza.44 The map below illustrates the locations where the study was conducted. 

In the north, Bondo is characterized by rolling plains, sandy soils, grasslands, hills, four islands, and 67 fish landing beaches. Fishing is a major component of the district's local economy. Kisumu district, the location of the capital of the province and the nation's third largest city, is south and east of Bondo district. It is a vibrant hub not just of the province, but indeed of the entire region, and it lies not only on a major highway and railway line leading to Uganda and West and to South Sudan and North, but also, along with Uganda and Tanzania, on the banks of Lake Victoria. Thus, it serves as a major port city that facilitates trade between Kenya and the two other East African countries, and as a pass through for trade from landlocked African countries to the coast of the Indian Ocean in southeast Kenya. Kisumu has about 50 miles of shoreline and 13 fish landing beaches. Additionally, many factories and trading centers are based here. (See figure 5.)

To Kisumu's south and east lies Nyando. It is characterized by three distinct geographical features: the Nandi Hills, shared with the Kalenjin and Nandi peoples; Nyabondo Plateau; and the Kano plains. On the plateau, because of the bounty of natural stone, the key industries are brick and concrete making; on the plains, rice farming is an important industry. (See figure 6 and figure 7.) 

Finally, on the far end of the lake, directly south and across the water from Bondo is Homa Bay. Its landscape is characterized by lowlands along the lakeshore and upland plateaus, and also has several beaches and small islands near its coastline. Distinctive features of Homa Bay are the hump-backed hills that line the landscape as one approaches the town as well as several rivers that crisscross the district.45 (See figure 8.)

Study Methods

Qualitative Fieldwork

My primary fieldwork was conducted between December 2005 and August 2006 and included 74 individual and focus group interviews among 185 youth (15-29 years old), middle-aged adults (35-55 years old), and older adults (56-93 years old). Since then, I have returned at least once a year to the province, though not in a research capacity. To begin my study, I randomly selected 15 secondary (high) and vocational schools in the locations of interest in Nyanza province from the Kenya Education Directory, an annually updated, complete listing of all the schools in Kenya, and then proceeded to each one.46 The schools were varying combinations of public and private, rural and urban, provincial and district, boarding and day. Rural schools often required a few hours of travel by road followed by several miles of walking. The experience of getting to these schools highlighted the physical isolation of many of them from even the nearest town, as well as the surprising reach of globalization and globalized messages of consumption, modernity, and HIV/AIDS. Selecting schools at random allowed for several starting points through which to conduct snowball sampling into different communities around the province. I conducted most community interviews in homes immediately surrounding the schools to capture young people who were not in school, as well as middle-aged adults and older adults. Additionally, respondents were recruited as individuals and in groups from community settings such as market places, villages, hospitals, community centers, and fish-landing beaches, and during social events where locals were gathered. Here I was reliant on friends of friends who served as local contacts, friendships and acquaintances developed in the course of my time in Nyanza, and the kind of boldness to simply start up conversations with strangers often required of field researchers. This approach allowed for data gathering from a combination of formal and noninstitutionalized settings. I also conducted interviews with 20 key informants, including government officials, local researchers, teachers, and community leaders. I asked study respondents about general issues facing the province (an icebreaker question initiating interviews), questions relating to young people's transitions to adulthood (about relationship formation, education, employment, and financial access), and questions about whether they felt HIV/AIDS was a problem in their community. Research material also included regularly written field notes documenting ethnographic observations and aspects of the culture and people's everyday lives not captured by formal taped interviews. The study protocol was approved by the University of Chicago's social sciences institutional review board, the Ministry of Education, Republic of Kenya, and the Nyanza district and school officials overseeing the sites of study.

My two Luo research assistants were able guides and companions as I made my way through fieldwork: Marian, who had recently graduated high school, accompanied and assisted me during most of my fieldwork; and Rose, who had a master's in sociology, provided assistance in the final month of fieldwork while Marian transitioned to university to pursue her bachelor's degree in education. I benefited from many interesting conversations with them over the months of fieldwork as we talked about what we had seen or heard in our travels. Conversations with Marian, who was a peer in many ways of those I was interviewing, were especially beneficial. She corrected some of my assumptions, and noted things about different settings that, as a nonlocal, I would not have noted. For example, we realized that we had different ways of assessing whether a household was wealthy when, after an interview in a hut, she noted that the household was relatively wealthy because they had a cow in the yard, unlike other households in the area.

Interviews were conducted in English and Kiswahili (the national languages), which, as a Kenyan native, I was fluent in, as well as in dhoLuo, the local ethnic group language, which I learned in the course of fieldwork. Along with Marian and Rose, I transcribed taped interviews and, where needed, translated them into English. I then conducted all coding and analyses of interview transcripts and field notes recording ethnographic observations. Both my data collection and analysis were guided by grounded theory; emergent themes and theories from early parts of the fieldwork shaped subsequent interviews, and a variety of interview themes were pursued until data and theoretical saturation was reached.47 In the chapters to follow, where respondents spoke in English, I try as much as possible to preserve their original phrasings and grammar to avoid making them sound like they speak Queen's English, and I provide explanation or clarification where needed. All names of respondents and schools are pseudonyms to preserve privacy and anonymity.48 For the same reason, because towns and villages are small and because many settings were characterized by thick social networks of friends, relatives, and acquaintances (as I quickly came to realize), I do not explicitly name the specific locations I visited.

Survey Data

I complemented qualitative fieldwork with quantitative survey analysis using data from the 2003 and 2008-9 Kenya Demographic and Health Surveys (KDHS 2004 and KDHS 2010). The KDHS surveys are nationally representative, household-based surveys of women (15-49) and men (15-54) that have been conducted every five years since 1989. The last two surveys, in 2003 and 2008-9, included HIV testing, with fairly high response rates for men and women.49 All statistical analyses in this book are based on those who were tested for HIV. I draw on KDHS 2003 for historical trends especially in chapter 3, and on KDHS 2010 for contemporary trends throughout the rest of the book. While the survey data are robust for estimating national trends and carrying out statistical analyses on the national data, a key limitation in their use for this study is the small sample sizes of Luo-Nyanza respondents (respondents of Luo ethnicity residing in Nyanza province), which limit similar group-specific analyses. Because of this limitation, my primary use of the data for the Luo-Nyanza population in this book is descriptive, presenting population-weighted trends of a variety of factors such as HIV prevalence, education, and employment. I also provide sample sizes and confidence intervals of these estimates.

Mixing Methods

Grounded theory also shaped my back-and-forth movement between survey and fieldwork data both during data collection and in subsequent data analysis and writing. Emergent codes in the qualitative material were explored in the quantitative data and vice versa. For example, the gender disparity in HIV rates in the survey among Luo-Nyanza youth prompted focus groups and individual interviews that concentrated on young men's relationships with same-aged partners (or lack thereof as it turned out, which is discussed chapter 4). Interview reports of an increasing number of relationships of men just out of school prompted closer survey analysis of how young men's HIV rates increased each year as they approached their late 20s. And, the nonlinear nature of the link between educational attainment and HIV rates revealed in survey analysis provoked a deeper qualitative exploration of what was going on in school (chapter 5).

My main assumption in using this kind of mixed-method analysis was that while appreciating that different logics of inquiry underlie the interviews driven by structured categories that produce survey data and the unstructured fieldwork that allows categories to emerge, they are nonetheless accessing different aspects of the same social reality inhabited by the people of Luo-Nyanza. In particular, characterizing and understanding the HIV risk environment in which young people were navigating their relationships required attention to three things: the numbers (for example, who was infected, by how much); the lived experiences of those embodying those numbers (for example, the logics guiding relationship decision-making); and the social-structural processes both producing those numbers and shaping those lived experiences, especially in institutional contexts such as schools and labor markets, which involve so many young people.

Reflexivity and the Marginal Woman

For much of my fieldwork, though I traveled and stayed in different areas throughout the province for varying lengths of time, my primary base was in Maseno, a small town in Kisumu district, located 40 minutes away from the province capital, Kisumu. (I was staying with paternal relatives who lived near the main road.) With easy access to public transport, fairly reliable electricity, and space to write, their place was ideal for my work. In several senses, I was Robert Park's "marginal (wo)man," belonging, yet not quite belonging, understanding, yet not quite understanding. I was familiar with the mud-floor houses and kitchens, the corrugated iron and thatched roofs, walking through the hills on narrow pathways, the slower pace of life, timing my travel to avoid torrential downpours, negotiating public transport, and the endless waiting for public officials. There was no learning or assimilating there. I was unfamiliar with much else: the language spoken by the ethnic group (which I had to learn), the daily mundane logics underlying how the Luo people lived life and approached situations, and the varying ways in which I was perceived as I traveled throughout the province. I engaged in participant observation and took regular field notes. I talked with local informants, getting to know them and learning their world through the words they said, the things they did, and the places they took me. I wrote field notes on everything I saw, from my daily transportation experiences, to descriptions of the landscape, my own feelings about particular places and people, events that took place, and so on. I talked to many people I encountered, such as a lady on the beach selling fish and the bike taxi men who transported me long distances to schools and other locations. In community interviews, I was highly dependent on informant networks-who they knew and which groups they could convene at short notice.

Fieldwork was characterized by many moments of mutual recognition and strangeness. I was perceived as a strange anomaly in a context where many young women drop out of school to get married, since I was an unmarried, childless young Kenyan woman studying for a PhD in America, the land of many people's dreams. But I looked like them, sometimes sounded like them, dressed like them, and shared their nationality and heritage. Yet perhaps the most powerful personal moments of recognition throughout fieldwork and after were when I observed the similarities between the consuming young women who are the subjects of this book and those I observed while I was writing and later teaching as a graduate student at the University of Chicago and then as an assistant professor at the University of Colorado in Boulder, several thousand miles, but not worlds, away. It was just that in Africa, young women's lives were more likely to end in tragedy. Far from an impersonal and purely academic account, then, this book is a study about the young women whose dangerous transitions I might have had to traverse except for the strange turns my life has taken.

Chapter 2 examines the widespread nature of transactional sex relationships, both in Africa as well as the United States. Specifically, the chapter illustrates the centrality of consumption to intimate partnerships in both settings, explores how consuming women are cultivated and produced, and examines similarities in the gendered entanglement of love, money, and consumption among women in vastly different places and historical time periods. In doing so, I show not only why women living in the wealthiest African households have the highest rates of HIV, but also how this entanglement places young African women at disproportionate risk for acquiring HIV.

Chapter 3 begins with a discussion of the density of death that resulted from living in a high HIV-prevalence environment, and how the Luo of Nyanza province navigated this social reality. I then describe the study's historical and cultural context to set the context for the contemporary transitions to adulthood that are explored in subsequent chapters. Specifically, I describe how young people in this setting historically came of age, their relationship-formation processes, the changing socioeconomy among the Luo ethnic group over the twentieth century, and consequent shifts in attitudes toward mass education, salaried employment, norms of beauty, money, and consumption.

In chapter 4, I show how intimate relationships and communities enable and motivate girls' transformation into consuming women. The chapter examines how Luo youth characterized their relationships (for example, "relationships for sex" and "relationships for education"), how they initiated relationships and found partners, and their motivations underlying partner selection. I connect these choices to the HIV trends, showing the epidemiological riskiness of young women's partnerships. Specifically, I show how these characterizations and motivations combined to make the selection of "safe" same-aged partners relatively less attractive (at least in the short term) than the selection of "dangerous" older partners. I show how money came to be central to intimate relationship formation, and I link this to historical changes in the socioeconomy, ideals of beauty, and consumption practices documented in chapter 2. Finally, I explore the implications of relationship logics in this setting in relation to key HIV-prevention strategies (abstinence, condom use, faithfulness, and HIV testing).

Chapter 5 shows how girls are produced as consuming women in the institutional context of school by exploring the complicated role of school in young women's lives. I show how schools, charged with constructing modern subjects, also contribute to the construction of gendered processes that place young women at risk. I begin with a discussion of the paradox of educated women's HIV risk and the nonlinear relationship between education and HIV. I then describe the rise of mass education in Africa, and Kenya in particular, before discussing the contemporary pursuit of education, including the rising constraints to access, associated expenses, and subsequent disillusionment. The rest of the chapter explores how school complicates young women's attempts to stay HIV free. Specifically, I explore the gendered construction of "needs" and its ties to consumption and modernity. I show how social-structural processes in school (for example, peer groups and the structure of the school calendar) both entrench these constructions and inadvertently facilitate transactional relationships, even while ostensibly seeking to regulate schoolgirl behavior. I discuss the dueling normative systems girls are choosing among, one urging them to wait to consume, and the other urging them to combine consumption and relationships with education. I conclude the chapter with case studies of young women's pursuit of education, as well as with reflections on the implications of the chapter findings for school-going girls in other African settings.

In chapter 6, I explore how gendered labor markets exacerbate the challenges young women face once they leave school, having been produced as consuming subjects. I describe their experiences as they confront a challenging economic environment in which their great needs meet limited means. I describe the kinds of jobs young women take up, the disillusionment faced in dealing with unemployment and itinerant work, and the reduction in financial support from parents and guardians on leaving school. In particular, I show how youth transitioning out of school confront a gendered labor market, gendered in access to jobs and in types of jobs and, as a result, gendered in access to and amount of income. I show how this is consequential for relationship formation and HIV risk for young women, and why women who work are at particular risk for HIV. In the second part of the chapter, I bring together many of the themes discussed in this book by engaging in a close study of fishing, a major lucrative industry in this area. I explore how the ecostructural context of Lake Victoria mapped onto the gendered fishing economy and combined to shape young women's relationships with fishermen and their HIV risk. Specifically, I illustrate how environmental changes in Lake Victoria and the surrounding area conditioned the sexual relationships and sexual mixing patterns of fisherfolk and young women. I show how this was exacerbated by a gendered fishing economy where only men fished and had access to daily income, in contrast to many locals in surrounding lakeside communities who lived on less than a dollar a day.

In the final chapter of the book, I propose a gendered and life-course perspective to HIV prevention, and apply the findings from this book to a consideration of individual, institutional, and ecostructural HIV-prevention strategies that might be usefully combined to reduce young African women's HIV risk, both among the Luo and in other parts of Africa.

The epilogue reflects on the current push to scale up antiretroviral therapy across sub-Saharan Africa, and considers both its promise and its limits in preventing HIV among young African women.